There has been considerable interest in developing vaccines to treat prostate cancer, the most common malignancy in American men. Nevertheless, efforts to translate vaccines developed in transplantable murine tumor models to treat human prostate cancer have only achieved limited success. During the initial funding period of this project we utilized a novel transgenic mouse model to demonstrate that the inability to achieve effective anti-prostate tumor immunity through vaccination is likely at least partially caused by the inactivation of tumor-reactive T cells through peripheral tolerization. To overcome T cell tolerance, we and others have targeted various costimulatory pathways to redirect the response of T cells encountering tolerogenic antigens. In particular, we found that enforced stimulation of CD134 (OX40) and CD137 (4-1BB) potently redirects CD8 T cells exposed to cognate soluble peptide to undergo massive clonal expansion and effector differentiation rather than anergy and deletion. Importantly, vaccines incorporating enforced OX40 and/or 4-1BB costimulation inhibit the growth of a variety of immunogenic murine tumors and are the basis for clinical trials to treat human cancers. We further tested the potential of enforced OX40/4-1BB dual costimulation to redirect T cell responses under highly stringent conditions where tolerizing epitopes derive from a widely and constitutively expressed parenchymal self-antigen. In contrast to soluble peptide, enforced dual costimulation boosted clonal expansion but surprisingly not effector differentiation when CD8 T cells encountered cognate parenchymally-derived self-antigen. CD4 T cells, which ironically possess a weaker tendency to undergo effector differentiation, expanded and differentiated into Th1 effectors in response to self-antigen plus enforced dual costimulation. Thus, specificity to self-antigen uncoupled effector differentiation from expansion exclusively in the CD8 T cell pool, and indicated that the response of CD8 T cells to enforced dual costimulation is strongly influenced by the APC presenting the tolerizing antigen. Nevertheless, when CD4 and CD8 T cells simultaneously encountered cognate parenchymal self-antigen during enforced dual costimulation, CD8 T cells were pushed to undergo effector differentiation. Thus, in the absence of CD4 T cell help enforced dual costimulation expands harmless self-reactive CD8 T cells. Notably, this expansion of harmless CD8 T cells is analogous to the decade-old but unexplained clinical observation that tumor-bearing individuals often harbor clonally expanded populations of anergic CD8 T cells specific for tissue/tumor- associated differentiation antigens. Our model thus represents a unique opportunity to understand and overcome this stubborn barrier to tumor immunity. The experimental aims of this revised competitive renewal application will elucidate the mechanism by which harmless CD8 T cells are expanded, and importantly how they can be pushed to express therapeutically useful effector functions in a clinically relevant prostate cancer model.